DARPA wants robots that help clean up after disasters

In the not-too distant future, putting out hazardous chemical fires, working on damaged nuclear reactors and other dangerous emergency response jobs may be handled by robots.

The Defense Department’s cutting-edge research and development shop is looking for next-generation machines capable of working in a human environment by doing things such as driving vehicles, climbing ladders, handling power tools and turning pipe valves.

To push the boundaries of robot technology for disaster relief operations, the Defense Advanced Research Projects Agency is launching its Robotics Challenge. The structure of the challenge, which begins in October, is similar to those the agency ran in recent years to develop autonomous ground vehicles: Robots developed by participating teams will compete against each other for a cash prize. In this case the winning team will be awarded $2 million.

Related coverage:

DARPA's cheetahbot: Fastest thing on two -- robot -- legs

DARPA plans smart, tough, fighting robots

A key thrust of the program is developing robots that can use available human tools in a disaster site, from vehicles to hand tools. DARPA also wants technologies that allow non-expert operators to control robots, lower operator workload, and permit effective operation in degraded, low bandwidth and intermittent communications environments.

The Robotics Challenge consists of three major events: a virtual disaster challenge and two disaster response challenges. Like DARPA’s previous contests for autonomous ground vehicles, the robots in this event will compete with each other. The current event will focus on disaster response scenarios in this sequence:

• Drive a utility vehicle at the site.
• Travel dismounted across rubble.
• Remove debris blocking an entryway.
• Open a door and enter a building.
• Climb an industrial ladder and move along an industrial walkway.
• Use a tool to break through a concrete panel.
• Locate and close a valve near a leaking pipe.
• Replace a component such as a cooling pump.

While the robots must be compatible with human operators, environments and tools, they are not required to have a humanoid form, DARPA officials said. But the robots must be able to get into, drive and navigate a standard, unmodified utility vehicle along a roadway.

For the other tests, robots will have to move across rubble; move along human-sized industrial ladders and walkways; detect and close leaking valves; handle power tools, and be able to locate, remove and replace a pump that a human would be able to physically carry and manipulate.

It is up to the participating teams to decide how autonomous their robots will be, DARPA officials said. Human operators will supervise the robots in all of the events, but for highly autonomous machines, this could mean only a few high-level commands via a relatively low data rate communications link. For less autonomous robots, more low-level commands and the need for more sensor data from the robot to the operator will require higher data rates.

As a part of the challenge, DARPA wants to make robot hardware and software development more accessible and to lower acquisition costs while increasing capability. The agency will create and provide Government Furnished Equipment (GFE) to some participants. Officially referred to as the GFE Platform, it consists of a man-sized robotic system with arms, legs, a torso and a head. The GFE Platform will allow teams without hardware expertise, or even hardware, to participate, agency officials said.

Along with the GFE Platform the agency will develop a GFE Simulator, an open-source, real-time, operator-interactive virtual test bed system. The simulator will run models of robots, robot components and field environments prior to physically validating them on hardware systems.

By using a widely available and affordable validation system to test software and hardware components, DARPA wants to create an environment where developers can quickly create and test new robot designs at minimal cost and with high confidence of success.